Documents · RAG
Chat with your documents, privately.
Drop in a file or folder; Bedrova embeds and retrieves locally, and answers with citations you can click. Nothing leaves your Mac.
What you can do
- Drop in files or a folder — PDF, DOCX, TXT, MD, HTML, CSV/JSON
- Local parsing — PDF, DOCX, Markdown, HTML, CSV and JSON, with OCR for scanned pages
- Hybrid search + rerank — the right context, packed into the prompt
- Clickable citations — jump to the exact source
How it works
Embeddings + a local vector store.
- Chunked with stable source spans
- Local embeddings (Nomic, BGE, Qwen3…)
- Full-doc-in-context when it fits
- Rebuild / repair / delete any index
See the documents docs.
Personal, not team
This is local document and folder chat. Durable, shared or agentic knowledge bases are Packwolf's job. Bedrova stays personal and on-device.
FAQ
Can I chat with my documents privately?
Yes. Bedrova parses, embeds and retrieves your files entirely on your Mac. Nothing is uploaded, and answers carry citations you can click straight back to the source.
What file types are supported?
PDF, DOCX, TXT, Markdown, HTML, CSV and JSON, parsed locally out of the box. Scanned pages go through on-device OCR. Heavier layout and table extraction is available as an optional advanced parser pack if you want it.
Is this a team knowledge base?
No. This is personal document chat that lives on your machine. Durable, shared or agentic knowledge bases are Packwolf's job.